Godlike-powerful ancient aliens have vanished from the galaxy for unknown reasons and left behind various artifacts. Millions of years later, lesser species had found these artifacts and discovered that those were hyperspace FTL engines. Strap it to a reactor and it will get you to Alpha Centauri in two days, the usual space opera stuff.

They were able to disassemble it and successfully build copies, yet they had failed to find any explanation for why the engine worked at all. Apparently, the precision of manufacturing and the used building materials affect the capabilities of the engine in various ways, but beyond that, there were only wild guesses and fruitless experiments. The only thing that everybody could agree upon is that apparently, those artifacts were intentionally designed for the purpose of being found and replicated by the less advanced species (No sinister secret motives behind that decision in the setting - a Kardashev type IV civilization's reasoning is beyond the comprehension of those who barely crawled out of Type I classification, same as for why did they decide to bail).

So is this setup plausible? Can a Clarketech piece of technology be impervious to scientific analysis due to there being too large a gulf of missing scientific knowledge between the builders and replicators, while the replicators still being able to make their own functional copies of the devices even though they have no idea how or why they work?

While this is tagged "science fiction" it's a science fiction that more aligned with stuff like "The Expanse" rather than "Star Trek".

  • $\begingroup$ The most obvious answer is simply that the replicators have enough technology to do a very detailed scan of the artifact, and then replicate it atom by atom. No understanding needed. $\endgroup$ – Starsong67 Nov 26 '20 at 9:10
  • 3
    $\begingroup$ It might be a good idea to use biotechnology for some parts of the engine. Perhaps some of the more complicated logic or material properties could be outsourced to a microorganism which can be easily cultivated even when one does not understand what it actually does and how it does it. Like some green organic goo which somehow generates antigraviton particles and grows if you feed it with simple carbohydrates. $\endgroup$ – Philipp Nov 26 '20 at 15:02
  • $\begingroup$ This is almost how hyperdrives are in the star wars universe; some hyper-advanced civilization left behind not just the hyperdrives themselves, but also plans for how to build them, but no one in the in-universe-present knows how they work. $\endgroup$ – Hearth Nov 26 '20 at 18:26
  • 1
    $\begingroup$ Contact the movie used this. We received a bunch of plans to build the (enormously expensive, one-of-a-kind) device, which was then destroyed because of religion, so it is revealed that an (enormously rich, one-of-a-kind and dying) benefactor built another one. AFAI Recall, no one actually understood for sure how it worked, or what it did. $\endgroup$ – CGCampbell Nov 27 '20 at 13:19
  • 1
    $\begingroup$ @MarkRogers The Human colony on We Made It in Nivens' Known Space universe purchased the Quantum Hyperdrive Shunt manual from the Outsiders. It facilitates rapid construction and operation of FTL drives, much to the dismay of the Kzinti. The Quantum II Hyperdrive Shunt was purchased (again from the Outsiders) as an operating device by the Puppeteers and gifted to Humans and Kzinti alike in return for crewing a second Ringworld expedition. Now as to why an Outsider ship pursuing a Starseed zigged when it should have zagged and happened to pass by We Made It instead of Kzinhome... ;) $\endgroup$ – Eight-Bit Guru Nov 28 '20 at 14:22

22 Answers 22


Yes, the drive is NOT what enables Hyperspace travel.

The "Hyperspace Drives" your people are discovering and duplicating are merely the activators, the "keys", to the actual mechanism that is not available for scrutiny. Maybe it is in an alternate parallel dimension, maybe it is something programmed into the very fabric of space.

As a current real-world analogy:
A 10th-century tinkerer would be quite capable of building a perfectly working light switch, if they discovered a couple of installed and functional working models.

They would be able to duplicate the copper contacts exactly, the steel casing with a different but good enough substitute, the insulation with a lot of experimenting. They could even duplicate, with effort, the screws and fittings needed to mount the thing to those mysterious copper wires in the wall. Or discover that the mounting is optional, just a good idea for safety.

But would they be able to understand, or duplicate the LED light that is mounted in the ceiling, and controlled by that switch? Even with full access to the lights, too? And how about just why those copper wires in the wall seem capable of making small lightning and heat, even death, when touched?

Your hyperdrives are the switches, which can be duplicated. Not the LED light which shines, and definitely not the power station and distribution network that delivers the power.

The Ancients put the dangerous side of Hyperdrive technology out of your reach, but made practical access to it possible.

I chose LED ceiling lighting for my example as there is no ways a medieval alchemist/engineer could figure them out or duplicate them. With enough incandescent lightbulbs, a large team of willing volunteers, and enough time, I think that having such things to play with might be able to jumpstart a smart and diligent researcher to Edison's lightbulb levels of understanding. But a low voltage, non-heat source of light is several knowledge quantums further down the queue. We want the hyperdrive's actual mechanism to be similarly several quantums of knowledge removed.

  • 14
    $\begingroup$ @John But to duplicate a phone, you need tech VERY SIMILAR to what is needed to build a mobile phone network. I wanted an example where the activation means was very easily copied, using a vastly lower tech level than the actual product. $\endgroup$ – PcMan Nov 26 '20 at 20:12
  • 4
    $\begingroup$ @John I see where you're going, but a satellite is not necessary to make a satellite phone work anymore than cars need gas stations. Transceivers are the important tech. Tech needed to build both the phone and the network. If you can build a sat phone, you can figure out a terrestrial network. A better analogy might be a web browser. It can communicate with the network, and the network protocol can be reverse engineered, but itself contains no smarts. That's all on the server. $\endgroup$ – Schwern Nov 27 '20 at 3:30
  • 3
    $\begingroup$ Hey, coincidentally this answer meshes very well with the state of the FTL in my setting - the ancients basically rewrote the laws of physics to allow FTL to be possible, and it's perfectly possible that it is sustained artificially. $\endgroup$ – Darth Biomech Nov 27 '20 at 3:41
  • 12
    $\begingroup$ @John I think you're massively misjudging the relative complexity of rocketry and semiconductors. If you can make fireworks and navigate the seas you have everything you fundamentally need to put something in orbit. Integrated circuits are basically black magic, and even if you slice one up very carefully and look at it with a very good microscope, almost impossible to practically copy, let alone understand. $\endgroup$ – Useless Nov 27 '20 at 11:14
  • 9
    $\begingroup$ An actually genuine real-world analogy: Tribesmen in Afghanistan and the surrounding regions are known to be able to replicate captured AK-47 rifles down to the individual tool-marks and wear-and-tear scratches. They don't bother to learn which tool marks are actually necessary for functionality (like the rifling in the barrel), so replicate everything. The final gun works just as well as the original. $\endgroup$ – EvoGamer Nov 27 '20 at 15:57


Take a wound stator DC motor,

If you can identify steel, iron, and copper in a motor you find (or are given) then you can build a mechanical copy of it. If you build your copy precisely enough, then applying an appropriate voltage (with enough current) to the power connections will make it run. If you copied it precisely enough, then the copy will run - even if you don't know about magnets or electromagnetism. Power can come from batteries, which is a different technology from motors or generators.

The trick is, of course, figuring out that it takes current from a battery to make it run. You need at least some insight into it to make that leap. If you have batteries as a potential power source, then you will probably recognize the copper as a conductor, and think to try connecting a battery to the motor.

For comparison, look at AC induction motors.

They are constructed of the same materials, but require a very different power source. If you give an AC induction motor to a group that has batteries and wires, and they manage to copy the motor, then they are going to have a difficult time getting it to operate. They might find that applying current to the motor causes it to make a partial turn (some few degrees,) but they won't immediately be able to make it rotate continuously. With some study, they might come to the conclusion that alternatively applying power to the coils will make it rotate and from there work out how to power it - but that's going to take some time.

In either case, the operation of the motor will provide some clues as to how it works. In both cases it should be possible for a determined group to make a functioning copy.

A really good, long lasting and reliable copy will require learning more about the materials used. The bearings, for example, need a different steel alloy than the housing. That won't be obvious at the start, but testing the copies will show parts that wear out faster.

You don't have to understand electromagnetism to make a copy of a motor, though you will have to know something about metals and electricity.

For your Clarketech aliens, you posit that they made the machines such that they were easily copied. That means using techniques that rely on the material properties and shape to do the job.

Most things these days are built with some electronics to control processes and make them more efficient and reliable. Not everything is done that way, though.

If you have a microwave oven, you may have noticed that it "thumps" when the power goes on. The reason for that is that the actual microwave generating part is relatively primitive - it depends in great part on the properties and shapes of the materials used rather than using power semiconductors.

The "business end" of a microwave oven is a cavity magnetron. This is a microwave generator that works by "blowing" an electron stream past a bunch of reflective cavities. It works sort of like a whistle or flute - the moving electrons cause broadband electromagnetic waves, and the cavites resonate at a particular frequency. This makes electromagnetic waves at a frequency determined by the size and shape of the cavities.

You could build a cavity magnetron by copying an existing one. If you copied it accurately enough and applied power to it correctly, then it would generate microwaves - even if your knowledge of electricity is limited to batteries and conductors.

Similarly, the transformer powering the magnetron is "primitive." It is the source of the "thump."

The transformer in a microwave oven is built in a particular way that limits the power to the magnetron. High power electronics to limit the current to the magnetron are expensive, and have only in the last ten years or so reached the point where microwave ovens use electronics.

The old transformers were built to limit the current to the magnetron by being intentionally not very good transformers. In an ideal transformer, there's no current induced in the core so all the power is transferred from the primary side to the secondary side.

A microwave oven transformer core is built so that drawing too much current generates current in the core. Since part of the power goes into the core, the output current drops.

This is partially the choice of materials in the core, but it is mostly due to the way the core is built. Most transformer cores for high power are laminated cores - they are built with layers of thin metal sheets (laminates) that are electrically insulated from one another. In a microwave over transformer core, the layers are deliberately shorted together so as to allow the current induced in each layer to reach other layers.

So, the mechanical form and the materials used are primary features in a microwave oven.

You could replicate at least those two parts of a microwave oven, and get something functional (though probably inefficient and dangerous) just by replicating the shape and using materials matched as well as possible.

Assuming really advanced aliens intentionally making a device intended to be copied, I can imagine them making a deceptively simple machine that does seemingly magical things - but only if copied precisely enough.

  • 3
    $\begingroup$ Wow, that was a really articulate answer, you explained something complex with a couple of simple analogies - which were also about topics I am ignorant of - and now my overall level of knowledge has increased. Thank you! $\endgroup$ – Adam Menhennett Nov 26 '20 at 14:21
  • $\begingroup$ I am not sure if that is just an unconfirmed anecdote, but I read somewhere that the electric generator was actually invented before the electric motor, and that it was discovered that most generators also work as motors when someone accidentally connected one generator to another and the second started spinning when they spun the first. $\endgroup$ – Philipp Nov 26 '20 at 14:38
  • $\begingroup$ @Philipp: According to this, motors were invented slightly before generators, but it was suspected before hand that a motor would work as a generator. $\endgroup$ – JRE Nov 26 '20 at 14:45
  • $\begingroup$ Your "cavity magnetron" link is actually about induction motors. $\endgroup$ – user253751 Nov 26 '20 at 17:30
  • $\begingroup$ @user253751: Thanks for the heads up. I fixed the link. $\endgroup$ – JRE Nov 26 '20 at 17:54

Potentially, as long as:

The artefacts are capable of self replication

If an elder race is specifically trying to gift technology to a younger one they wouldn’t risk misunderstandings or confusion getting in the way. A much simpler and easier thing to do is make a high tech 3D printer that the younger race can use to print all the components for another 3D printer, along with an IKEA style guide on how to fit them together.

That way they can make the tech as mind boggling as they like, build in whatever safeguards they need to, and never risk the younger race actually understanding what they’re doing. As far as the young race is concerned they push the button, feed the magic box the right ingredients and get another magic box, or a magic cancer-curing wand, or a magic laser cannon.

The only comprehension required is how to operate the machines, not how the machines operate.

  • 1
    $\begingroup$ Like operating a car without any knowledge of gearboxes, fuel or transmissions. Just do what you've learned, put fuel in when indicated and get it to a shop when something goes bang. $\endgroup$ – Trioxidane Nov 26 '20 at 14:45
  • 7
    $\begingroup$ This sounds familiar... $\endgroup$ – Matthew Nov 26 '20 at 15:09
  • 5
    $\begingroup$ @Trioxidane: more like owning a stable of horses. Put in food and they do things for you. Put two of them together and eventually you’ll end up with more horses. $\endgroup$ – Joe Bloggs Nov 26 '20 at 21:28
  • $\begingroup$ @Matthew: Hah! Pretty much exactly that. $\endgroup$ – Joe Bloggs Nov 26 '20 at 21:28
  • 1
    $\begingroup$ Then again, IKEA assembly guides are examples of Clarketech :) $\endgroup$ – Hagen von Eitzen Nov 28 '20 at 12:44

Yes - and we've been doing it for a long time

As a very simple modern day analogue, consider the forging of steel. There are many levels to understanding, but at a very basic sense when we began forging steel it was a happy accident that led to a significant mythos of what you had to put in the iron to make it stronger (lots of "interesting" sources of carbon exist). Hundreds of years later we start to understand the chemistry of why it works, and can then develop better methods of manufacture, which leads to stronger steels again. Decades later again, we start to understand the quantum physics at the core of the chemistry and can then make even more specific steels to (in SpaceX's case) even potentially survive the heat of re-entry.

Could modern day metallurgy write instructions a dark ages blacksmith could follow? It's certainly feasible in the realm of sci-fi. Quality would be poor, and you'd have to introduce a ton of testing stages in terms available at the time, but it's feasible in a story!


Maybe this is obvious from your use of the word "Clarketech," but FWIW, this is exactly how magic spells work in most systems, right? The spellcaster knows how to reliably reproduce the effect of "wingardium leviosa" (namely "repeat these nonsense words in this order"), but doesn't have a clear picture of why it should work.

If your aliens are able to control enough of (perceived) reality, then you can posit arbitrarily symbolic "devices."

Our scientists have discovered that if you place a ham sandwich in a vacuum chamber of volume approximately 0.5 liter, it becomes a FTL drive.

This might be because the aliens have designed and/or modified the parameters of our physical Universe to produce this specific effect. Or, it might be because the aliens already control our perceptions and so there's no observable distinction between "this sandwich is an FTL drive" and "you perceive that there is a sandwich here and also that it is an FTL drive."

Alternatively (maybe more hard-sciencey), I like PcMan's answer which boils down to that the ham sandwich is merely a license key. Perhaps the aliens are constantly observing our physical Universe, and when they observe a ham sandwich in this configuration, they interpret it as a request for an FTL drive to be constructed there. Perhaps omnipresent nanobots are involved. (The Three-Body Problem uses basically this plot device. Maybe The Diamond Age too.)

Both of these suggestions are cheating in the sense that they involve giving the aliens some degree of interactive control over the physical vicinity of the ham sandwich. (Even if the aliens are long-dead, they've simply delegated control to their still-extant nanobots and/or still-extant parameters-of-the-Universe.) If the parameters of the problem rule out such interactive solutions, then personally I'm stumped.

  • $\begingroup$ Where does the Dark Forest trilogy use the concept of omnipresent nanobots in this way? Are you talking about the Three Body Problem/Remembrance Of Earth Past trilogy? $\endgroup$ – Aron Nov 29 '20 at 16:08
  • $\begingroup$ @Aron: Yes; I muffed the name and will fix. From Wikipedia's summary of book 1: "[The Trisolarans] create eleven-dimensional supercomputers called sophons which[...] occupy the volume of a proton. Two of these sophons have been laboriously manufactured and sent to Earth, having the power to cause hallucinations, spy on any corner of the Earth, transmit the information gathered to Trisolaris[,] disrupt all of Earth's particle accelerators[,] fabricate visual miracles and other hallucinations on a massive scale [...]" They're not called "nanobots" in-universe, but fill the same narrative niche. $\endgroup$ – Quuxplusone Nov 29 '20 at 17:02

You don't need alien artifacts - that's how the things currently are.

The other answers mention DC motor, or light switch, that can be copied without understanding. But is our "understanding" any better? We do know that the motor works because Lorentz force acts on charges moving in magnetic fields, but why does it? We can bring Maxwell equations and special relativity and whatnot, but that only shifts this question "why" one layer further. Eventually, you always hit the layer where you cannot answer "why" in any way other than "that's how the things are". Watch Richard Feynman making the same point.

And that's in physics, where we at least believe to have discovered a finite set of fundamental "how" properties that are supposed to be able to explain all our "why" questions. But even many questions that we are supposed, in principle, to be able to answer from this first principles, in practice have no better answer than "because that's how it is." Why is a particular alloy superconductive at a particular temperature? Why is a particular chemical compound efficient against a particular disease? Why does a particular virus cause cytokine storm and others don't? Why do neural networks learn? Eventually, all our answers to why questions are just a way to organize some pretty modest part of our knowledge about how things are.

So, your FTL devices clearly will have to add something new to our very bottom layer of our how things are knowledge. There will be five fundamental interactions: gravity, electromagnetic, weak, strong and this mumbo-jumbo that makes FTL devices fly. As we tweak the devices and see what happens, we might find more about the mumbo-jumbo, but that knowledge may simply not crystallize into any compactly formulated set of simpler principles like Standard model or General relativity. Or we might actually find out such principles, but lack computational power to do anything meaningful with them - just as we lack computational power to design a molecule that will interact with Covid RNA but not human RNA.

As an concrete suggestion, mumbo-jumbo interacts with usual matter very weakly, but certain configurations of certain materials ("magic crystals") produce resonances that can amplify the interaction and create disturbances in space-time. These resonances and disturbances are highly non-linear, so even as people understand the principles of interaction, an immense computational power is required to figure out, for a given configuration, which disturbances will it produce. Even more complicated is the solution of the inverse problem: design a configuration that will produce disturbances of the shape useful for FTL travel. A quantum computer consuming all the energy of the sun would require 1 million years to solve the problem. A Kardashev 3 civilization could afford that, but the humans can only replicate their configuration.

  • $\begingroup$ I think this solution fits the criteria very well. If a fifth, or combination of multiple, extra forces were at play, with sufficient complexity, it could produce phenomena for which we would be unable to create a theory. $\endgroup$ – user110866 Nov 27 '20 at 2:10
  • $\begingroup$ Well, the lore of my setting is that the ancients explicitly were stuck in a universe governed by physics as Einstein envisioned them, until they developed enough to say "no, we want FTL, now bend over" to the universe, just to drive the point home that a civilization billions of years older than you will be way more lovecraftian than just, say, having handguns that can blow up planets, or claim that they're superior to everybody else. $\endgroup$ – Darth Biomech Nov 27 '20 at 3:52
  • $\begingroup$ @DarthBiomech, well if you presume that they still didn't bend the laws of the Universe by sheer willpower, one solution would be that they decided to invest into titanic amounts of computations. I added a suggestion. $\endgroup$ – Kostya_I Nov 27 '20 at 12:22

I don't think it is plausible: an assembly is more than its individual parts.

Take a car engine: one can replicate all its components, but when they are put together without the knowledge on how to properly sync the valves with the piston the end result will be a CLANK BANG not a running engine.

Even worse if one takes something that has software or IC in it. One doesn't see the software in a ROM, but try charging a lithium battery without any software controller on the charger.

Once I was talking with the install engineer of an electron beam microscope produced in Europe: he was telling me that in some far East country somebody had tried to copy one of their machines to make a cheaper version without worrying of patents. Though they managed to make a fairly good physical copy of the whole thing, they weren't able to make it work and had to give up. Something similar happens also with car copies. And an E-beam or a deluxe car is nothing as complex as a Kardashiev IV product.

  • 1
    $\begingroup$ If you fail to take account of the relative positions of mechanically linked components, that's your own lookout really. $\endgroup$ – Separatrix Nov 26 '20 at 13:29
  • 2
    $\begingroup$ But neither cars nor electron microscopes are specifically designed to be easy to replicate without special knowledge. I don't know much about electron microscopes, but I am sure that it is possible to create a car engine in a much more "idiot-proof" design than that of an average modern car. $\endgroup$ – Philipp Nov 26 '20 at 14:53
  • $\begingroup$ @Philipp That's an interesting point, but let's counter it with why car engines aren't idiot-proof. (a) Government regulation requiring components for (e.g.) emissions control that are not necessary for the operation of the engine. (b) Engine components that enhance power (e.g. turbos) that are not necessary... (c) Engine components that improve diagnostics or operating efficiency (e.g., sensors) that are not necessary.... My point is, it's almost never desirable to design something that's idiot proof. It's therefore a bad assumption that Clarketech would be or even should be. $\endgroup$ – JBH Nov 26 '20 at 20:49
  • 1
    $\begingroup$ Ironically, a car is a bad analogy here. You can make a really badly-tuned engine that still runs. You don't even need anything fancy for fuelling - early cars just blew air over a little tub of gasoline. Getting modern performance is hard, but getting something which kind of works is easy. The reason it took until late-Victorian times isn't the concept, it's simply the quality of steel you need. As an engineer, I'm well aware everything I do could be copied - my protection is simply that it's not worth the time commercially for what I do. $\endgroup$ – Graham Nov 26 '20 at 22:39
  • 1
    $\begingroup$ @JBH But that's the whole premise of the question: The sufficiently advanced aliens designed the FTL engines with the explicit goal to be easy to replicate by less advanced intelligences. Being efficient, compact, fast, commercially competitive, maintainable or fulfill some intergalactic government regulation is not on their requirements list. Perhaps if the aliens tried, they could travel lightyears in seconds instead of days using a device which fits in their pocket instead of the size of a truck. But there is no way the lesser lifeforms could replicate that technology. $\endgroup$ – Philipp Nov 27 '20 at 8:55

Yes, why not?

I'm going to crib from The Crystal Spheres, Ask a Foolish Question and The Naked God. And The Last Question for good measure.

You say we have a K-IV civilization - something like the originator race of AAFQ:

Of the race that built him, the less said the better. They also Knew, and never said whether they found the knowledge pleasant. They built Answerer as a service to less-sophisticated races, and departed in a unique manner. Where they went, only Answerer knows. Because Answerer knows everything.

So, this K-IV race has godlike powers, as a K-IV race is wont to have. They decide to help somewhat the less-sophisticated races; but you cannot give matches to children, and FTL travel technology is potentially arming those children with RPGs.

The K-IV guys want to help the more backward races, but they don't want one of them, say, to enslave all the others.

And it would happen because, to have FTL, you need to have, say, adeledicnander generators, and if you have those, it would be trivial to reconfigure them into weapons of mass destruction.

So they cheat. They design and build machines that do something simple, much like a dynamo does. But while doing the only thing they actually can do, those machines also generate, as a secondary effect, specific spurious signals - just like the EM interference of a real dynamo. There is no reason they shouldn't, and there's no reason to investigate those - everybody knows they're just interferences.

Very far away, and yet not very far at all, in a dimension someone might call hyperspace, other machines lurk, and detect those interferences. When they establish that a viable "FTL engine" has been activated, they activate the real generators and pluck the engine and whatever it is attached to out of normal space, to reintegrate it in the appropriate location of the space-time continuum. This is

what the Naked God does in the third book of The Reality Dysfunction series: it can project massive wormholes on a galactic scale.

At any moment, at the K-IV people's whim, the FTL "engines" could stop working, or stop working for someone specific, or malfunction in any way. Because they're not FTL engines at all, they're just bells. You ring the bell, and if the correct note is struck, the bellboy comes and the service gets done - but it's not by any inherent power of the bell.

Being comparatively simple, it would be possible to replicate these "FTL engines" with little trouble, and not being FTL engines at all, they can't be reengineered or understood.

Actually, they could be just dynamos. A heavy dynamo with an alternated osmium-aluminum rotor connected to specially shaped coils. When operating, the device releases extremely weak gravitational waves, and inside it there are characteristic electromagnetic waves of exactly the same frequency. The source of both disturbances is pinpointed by the hyperspace controllers - using something like Iain M. Banks' Culture's effectors - that are able to analyze the device and "read" the current in the coils, then translate this information into jump parameters (say, every ampère of current in the coils is one light-year in the same direction as the dynamo axis).

Using the same trick, the hyperspace controllers can inject a current in a properly shaped coil, and "leak" information about e.g. nearby masses or FTL engine "wakes" or things like that.

So you have it - from the lesser races' point of view, a not too complex contraption made of rare earths magnets and both very dense and very light metals, rotating at a specific speed, opens a "portal" - then the current in the coils "drives" the ship through the portal at a "pseudo-speed" that depends on the ship's mass, the current in the coils, their diameter, the voltage, whether they're immersed in a magnetic field and so on.

Plot tools

Lots of experimentation and more and more outlandish theories would ensue, but the operating principle of the machines would remain mysterious. This does not mean that experimentation is useless; it leads to the discovery of the FTL "rules". Which can be as arbitrary as we need.

Rotate them slower than the threshold, and nothing at all happens. Increase the current or the coil area, and the distance covered changes - but that, while almost making sense, is of very little practical use. The engines might consume so little that they can reach any distance; but, after a jump that is never longer than, say, ten light-years, nothing can jump to or from the same volume of space (say, ten light-minute in radius) for some time, say four hours. Attempting to jump in a "depleted" or "hot" volume before it has "recovered" or "cooled down" means destruction of the engine (or maybe the ship?). This means that information speed, using relay couriers, is now one light-year per minute; ship speed is 2.5 light-years per hour. Also, a volume of space can be made impervious to FTL by having several FTL drones, ten light-minutes apart, hop around the whole volume. Precise knowledge of the drone schedule allows FTL travel in and out of the volume; jumping at random means almost surely smashing the engine into the "depletion shield".


Yes, consider a satellite phone.

You can have something you understand but cannot replicate, that's fairly easy, but there is only one believable ways to have something you can replicate but do not understand.

You only have part of the device, PCman hints at this. like having a cellphone is great but a cellphone does not work without the cellular network. A satellite phone does not work without satellites, An electric smelter does not work without an electrical supply. We have a lot of technology that relies on other technology that are not part of the same device. A new sat phone can link to an existing satellite network, but being able to build the phone does not tell you how ot build the satellite. The original might work by connecting to something we can't see/physically acess and thus cannot replicate.

You hyperdrive connects to the hyperdrive network, which is IN hyperspace and is all the stuff that makes it work, the "drive" is really just a connection device to the network not the things that make the network function. The network is what actually moves your ship, the "drive" is just an access device.

  • $\begingroup$ This was a clever solution to the problem and a good one for the OP to consider. The target society actually understands the object they duplicated just fine. What they don't understand is how the other part works. +1. I like this idea. $\endgroup$ – JBH Nov 26 '20 at 21:20
  • $\begingroup$ @JBH they may even understand how the rest of it works just not how to get the rest of it were it needs to be. "so there is a massive hyperspace shaper in hyperspace that moves the ship for us," "Yes" "but he transmitter uses a tiny hyperspace shaper so can't we scale that up?" "maybe but they won't do what we want unless we put hyperspace, and we don't have hte slightest clue how to permanently put something in hyperspace $\endgroup$ – John Nov 27 '20 at 4:53
  • 1
    $\begingroup$ Excellent point! The phone teaches us a lot about transmitting signals through hyperspace, but nothing at all about traveling through hyperspace, which uses a different set of rules. This is not unlike the fact that we can communicate at the speed of light today but the only way to do it with Mars is to wait for a satellite to slowly get to Mars because we don't know how to travel at the speed of light. Dang. I like this answer even more. It's too bad I can't up-vote it again. It's likely the most believable solution here. $\endgroup$ – JBH Nov 27 '20 at 5:01

We have no fundamental understanding of how Quantum Mechanics "works".
Those who 'understand' don't.
It 'just does'.

Albert Einstein violently disliked QM's "spooky action at a distance" - but experiments since his death have confirmed the reality in our reality of such "nonsensical" effects.

And yet, QM is arguably the most successful scientific theory of all time.
It's predictions are probabilistically precise and found to work (so far) with perfection.

With QM the 'how it works' and what it does are essentially orthogonal.
The effects may as well have been designed by aliens.
Now, there's a thought.



As another answer points out, smiths of the past knew how to make "good steel". They had no idea why it worked, or that what was really happening was alloying carbon with the steel, because the atom hadn't been discovered yet, never mind identifying "carbon" and "iron" as independent chemical elements. All the same they were able to work out procedures to reliably produce good steel.

For a more recent example, consider antidepressants. Evidence is pretty good that they work (even if exactly how well is up for debate). But we don't really know why they work. We have a good idea what the effects of taking them are, generally. But we don't even know which of those effects are therapeutic & which are a side effect, never mind how the therapeutic effects actually fix the problem.

In general, it's easier to prove some device works, than to explain how it works. Proving it works is just a matter of running a few well-designed studies - keeping notes, and crunching some numbers, basically. Explaining how it works might require principles we don't even know exist.

Thought experiment: Time traveller gives X-rays to the 1200's

Imagine giving someone from the 1200's an extremely detailed instruction manual, outlining step by step how to construct an X-ray machine (the medical device) completely from scratch using period-appropriate technology.

It explains everything they have to do, from how to build the tools they'll need to have in order to build the better tools needed to actually make the thing, to how get & refine the raw materials, to operating the device & developing the pictures. But not one word about any underlying theory.

They won't have any idea how any of it works. Physics hasn't even been invented yet, never mind nuclear radiation. But they can easily see it does, in fact, work - they x-ray a subject, they can see the bones, and can easily verify they really are seeing them by (for example) breaking some bones of a pig carcass.


Yes, but

It must be fundamentally designed to be copied by a civilisation with a given minimum technological capability.

This is the story of the first Krikkit starhip. They copied it from a "crashed starship" that rather than being the geniuine article was in fact a full blueprint for a starship carefully engineered such that they would be able to make their own.

For real technology, no chance. Even our own tech is a black box to the average user, never mind someone from a primitive culture. We also have a tendency towards DRM and the like, actively preventing copies if at all possible. There's no reason to believe that an advanced culture would want their technology copied unless they had an ulterior motive, as such their technological artifacts would not be copyable by anyone less advanced than themselves.

  • $\begingroup$ Out of curiosity, in the referenced story, did they not actually comprehend the technology they were using for space travel? The idea that you can duplicate it and use it but not comprehend it feels like an oxymoron. $\endgroup$ – JBH Nov 26 '20 at 20:51
  • $\begingroup$ @JBH they were an agrarian society before the encounter, specifically with no knowledge of the existance of anything beyond their own single planet and star. $\endgroup$ – Separatrix Nov 26 '20 at 20:56

If it's within our fabrication tolerances, of course we can.

To copy tech at a high level, there's really 2 parts to:

  • Scan it in. How accurately can we turn the atomic layout into a computer model?
  • Print it out. How accurately can we turn that model into a physical thing?

Scanning we can do really well if properly motivated. It's totally possible to:

And fabricating we can do pretty well too:

So long as the advanced technology we don't understand can be captured by our scanning tech, and reproduced by our printing tech, we can copy anything without needing to understanding it.

  • $\begingroup$ "David Blaauw, a professor of electrical and computer engineering at University of Michigan"... I wonder whether he was Gerrit Blaauw's son? More photos at blaauw.engin.umich.edu $\endgroup$ – Mark Morgan Lloyd Nov 26 '20 at 18:25
  • $\begingroup$ We still haven't been able to recreate an artificial human brain, for all our abilities to scan at the molecular level etc. $\endgroup$ – Slarty Nov 26 '20 at 18:38
  • $\begingroup$ Am I misunderstanding your answer? The question is, is it possible to have the ability to duplicate something without understanding it. I'm an electrical engineer and I've reverse engineered circuits. It's impossible to do so and not comprehend what you did. $\endgroup$ – JBH Nov 26 '20 at 20:55
  • 1
    $\begingroup$ @JBH I have half an electrical engineering degree (switched to software 2 years in). I can identify every part, print a pcb, etch it, solder it, read and write firmware byte by byte, and confirm that two circuit boards are identical, but for the life of me I cant explain how they work. Especially analog circuits. Totally over my head. But I can duplicate them. $\endgroup$ – Ash Nov 27 '20 at 0:52
  • $\begingroup$ @Ash You might not, but I can (in the case of microelectronics). That's the point. If the OP were only talking about an individual, the answer would be "absolutely yes." But the OP is talking about an entire species. There's always someone who can explain it... otherwise you couldn't perform the duplication in the first place. And even if it were possible that no one could... how long would it be before they figured it out? After all, the duplication tech exists, which means no technology or physics are in play that aren't within the target tech level. $\endgroup$ – JBH Nov 27 '20 at 0:57


Assumption: The target civilization is technologically capable of replicating the Clarketech.

Assumption: The duplication results in an object that is 100% identical to the original. In other words, you can't replace a miniature fusion generator the size of a coin with a battery that only lasts a micro-second and claim to have duplicated the object. Every component must be an exact duplicate of its original counterpart.

What you're suggesting is that every component of the Clarketech is manufacturable by the target civilization, but that something about the assembly of parts cannot be understood, despite the duplicate working perfectly. For example, one of the components may be a curved piece of aluminum (something easily manufactured by us today, and so a good example of the problem), which can be easily replicated by the target society — but its use in the assembly is a mystery.

I cannot find that situation believable. Maybe for the first couple of weeks, but (proverbially) every scientist on the planet would be working on understanding why that curved piece of aluminum had to be as it was. They'd be testing different shapes, different qualities of aluminum, etc. And they have a working assembly with which to test their ideas. The idea that such a civilization could create the mysterious object but not figure out what it does is, itself, incomprehensible. It's like suggesting it has the manufacturing chops to build and use a nuclear bomb, but can't comprehend nuclear physics. Basically, you're suggesting that despite having the ability to manufacture and, by extension, the ability to experiment and test, they can't comprehend it.

I'm fond of a quote from another literarily-gifted scientist:

The most exciting phrase to hear in science, the one that heralds new discoveries, is not “Eureka” but “That’s funny...” —Isaac Asimov

And you'd have every scientist on the planet scratching their heads and saying, "that's funny...."

While other respondents point out that the conditions of your question must change a bit to permit the lack of comprehension, taken at its word, your question's only practical answer is, "no, this can't be done."

  • 1
    $\begingroup$ I disagree. JRE's example of a cavity magnetron is a perfect counter example. Its not impossible to replicate, but without the knowledge and understanding of the physics behind it, you can't comprehend how it works. As long as the object's function depends on complex physical phenomenon that can't be observed (by the civilization in question), it will fit OPs criteron. $\endgroup$ – JS Lavertu Nov 27 '20 at 13:25
  • 1
    $\begingroup$ What if the machine does something that travels back through time and space? How would you measure it or find out? What if it does something in dimension X that we know nothing about, so can't see any measurable outcome besides the FTL? What if it is right under our noses, but simply beyond comprehension, like it changing the rules of the universe right before us. There is no way in telling what would happen and how we can start interpreting it. It might take millions of years to understand, if at all. $\endgroup$ – Trioxidane Nov 27 '20 at 14:51
  • $\begingroup$ @Trioxidane Good grief. It does something useful, doesn't it? A very measurable something. After that it's only a matter of time. And that's really the point. The aliens figured it out. Given time humans could (and would), too. Really, the OP's biggest problem is scoping how much time the humans must remain ignorant. The longer he needs, the less believable human ignorance would be. $\endgroup$ – JBH Nov 28 '20 at 5:00
  • $\begingroup$ And I feel it's like a 3 dimensional being telling a 2 dimensional one about powered flight. But the longer you need. $\endgroup$ – Trioxidane Nov 28 '20 at 9:31
  • $\begingroup$ @Trioxidane A 2D being would not be capable of replicating an engine capable of 3D flight. The point of my answer is that the conditions the OP set make it impossible to avoid comprehension - in the long term, if not the short term. $\endgroup$ – JBH Nov 29 '20 at 19:01


Software. Most of the "interesting" software in the world is already incomprehensible, at least for a single individual. It would be trivial for aliens to leave us software that does very useful things which we could easily replicate without having the faintest clue how it works (despite having an ability to completely debug and trace the code, because the scale of the code). But how do you build an FTL drive with software? That's easy too.


Creating a plasma is easy. You do it every time you turn on a fluorescent light. Heating a sample to a million degrees is also easy. We do that with high power lasers on a regular basis. Heating a plasma to a million degrees in a sustained fusion reaction is very, very hard. The problem is not that it requires materials we haven't invented yet, or energy sources or magnetic fields that we can't build. The problem is that plasma is a fluid and quite often behaves turbulently, and we don't have tractable equations to solve turbulent fluid dynamics in real time. It is barely a stretch to imagine an alien race who has left behind what they consider to be a "primitive" fusion reactor of the tokamak design we have been refining for 50+ years, along with the necessary software that makes it "just work". What the software would do is dynamically control the magnetic confinement field to maintain a stable, self-sustaining plasma. It also easy to imagine that this software could have completely transparent instructions that any programmer could observe in a debugger, yet have no idea how or why it works in the large picture (for instance, it could be implementing an enormously complex yet compact cellular automaton which just happens to solve the fluid dynamics problem of containing a plasma in a donut).

In a way, such a gift would be frustrating, because while it may be straightforward to simply copy the reactor numerous times, it might not be possible to scale it, depending on how the software works. Perhaps it gives an output of 5 GW, but if you want to build one that outputs 20 GW, the software fails spectacularly, even though the physical construction is more than up to the task. Or, the software could be so adaptive that it works over a large range of scales, depending on what result you want.

Structural Batteries/Computation

Right now, Tesla is transitioning from the "skateboard" design where their electric vehicles are built on a huge battery pack which forms the floor of the vehicle to one in which power storage is embedded into the frame of the vehicle itself. This is surely an embarrassingly crude first step in technology compared to what will likely exist on earth 100 years from now. In that not-too-far future, the idea of something "containing" a battery will seem as quaint as someone walking down the street with a boombox on their shoulder. But why stop at energy? Instead of running wires everywhere and centralizing computation, it seems just as likely that processing power will be similarly distributed. And so, we will have something vaguely similar to "programmable matter" that you see in so many SF stories. You don't need it to perform arbitrary shape-shifting feats to do something which is technically observable but practically inscrutable.

Whether we are talking about a starship engine or a food replicator, it is not too hard to imagine a level of technological craftsmanship which applies nanotechnology that would be familiar to engineers alive today combined with information technology that we can also recognize, but woven together at a level of complexity so far beyond what we can imagine that we must just take the informational gift at face value and thank our lucky stars. The analogy here is to take something like the OpenGL library, and gift it to a programmer from the 1950s. Even if you cross-compiled it for whatever hardware was available at the time, the sheer size of the code would outclass everything in existence and leave programmers completely baffled.

Now, if you included a thorough tutorial on 3D graphics primitives, then over time, they would come to understand how it all works. But if you just left them with binaries and a few programs to play with, they could likely learn the API to some extent, but being able to make significant modifications to it, or reverse-engineering the source code would be an amazing feat. That's just 70 years of software engineering. If aliens dropped code on us with 1000 years of engineering history, it would be exponentially more inscrutable.

Not only would it contain calculations we have not yet imagined, it may be so powerful that the computation substrate is able to reconfigure the molecular structure of the material in some way, similar to how EEPROMs and SSDs modify the physical material of their storage. This is important, because it may be that the most advanced technology cannot be manufactured with our equipment. Instead, we can only copy the crude starting point, which would include all the macroscopic structure. Then, when we activate the software, the device essentially finishes building itself on the nanoscale. This would render the final product obscure and opaque to us technologically, especially if we lacked the probes to scan the product at the finest level of detail.


In the crudest form, a bicycle today is not that different from a bicycle built 100 years ago. And yet, we have not wasted that 100 years of engineering. The improvements we make today are so subtle they would go unnoticed by that 100 year old bike builder, because they amount to removing a few grams of weight here and there. While there will surely be advances in metallurgy and materials sciences, I think by far the greatest advances will be in the information used to manufacture products. We can imagine future materials that are 10, maybe 100x stronger than what we have today; but it seems unlikely that we will get something 1 million x stronger. Chemistry has its limits, after all. But it is virtually guaranteed that the informational complexity of future tech will be well more than 1 million x what we have today. Giving us the raw information, without a tutorial, is equivalent to handing us magic. It can be done in a way that we can copy without comprehending. I can attest to this because I copied many a program from computer magazines as a child without having the faintest clue what the symbols meant. And yet, the magic proceeded to work anyway.


"Clarketech"... OK, if you insist. In one of the passages quoted in Clarke's "The Lost Worlds of 2001", somebody at a White House reception makes the old joke about humans being replicated by unskilled labour. Most things about what makes "a man a man for a' that" are still mysteries to the people who insist on making more of them, which I suggest answers OP's question.

  • $\begingroup$ This doesn't answer the question. $\endgroup$ – JS Lavertu Nov 27 '20 at 13:28

Perhaps they have a recipe rather than a blue print. Maybe they are told to synthesize a particularly complex range of chemicals, arrange them in a complex matrix of other compounds and are then told to add x, y and z in a specific sequence and at specified temperatures.

They follow the exact instructions and the mixture suddenly starts to self organize itself into more complex structures which then recombine into even more complex structures until eventually some form of artificial brain has been generated. They then have to feed the brain a huge mass of totally unintelligible data after which the brain is able to help them build the drive they need using similarly obscure techniques.

The scope within chemistry for complexity is huge given the correct starting point. https://www.reddit.com/r/chemicalreactiongifs/comments/4ia8ai/spiraling_demon_reaction/ https://www.youtube.com/watch?v=LL3kVtc-4vY


To continue on from JRE's example of the microwave oven, there is actually a thruster design that works on the same principle called a radio frequency resonant cavity thruster. Currently we can build them, but we aren't sure if they work (NASA Advanced Propulsion Physics Laboratory tested it and said it worked, but others haven't been able to replicate the results). However, they are in violation of several laws of physics, most notably the law of conservation of momentum.

So while this is not a FTL drive, it is a real life spacecraft thruster drive that we can build quite easily, but do not understand how it works - if it works at all.

Not only is it feasible, but it happens in real life.

  • 1
    $\begingroup$ Are you talking about the EmDrive? I thought it was debunked in that the "thrust" detected was caused by the induction in the cables that power the assembly. $\endgroup$ – Darth Biomech Nov 27 '20 at 4:01
  • $\begingroup$ @DarthBiomech That's a bit of an over-simplification, there has been a lot of different ideas about whether or not there are errors and if so where they come from. However enough scientists believe there's something there that despite these problems being raised one to two decades ago, research at NASA, DARPA, etc is still ongoing today. You can check the wiki article here: en.wikipedia.org/wiki/EmDrive It's a good real-life analogue to what OP is describing. $\endgroup$ – gbeeduljqa Nov 27 '20 at 4:31
  • $\begingroup$ You are right when you say "if it works at all". Not being replicated by other experiments is red flag for it doesn't look like it works. Besides if it did work the thrust will be microscopic. Satellites try use it, assuming it works, for manuevering very, very slowly. Not enough for a practical flying saucer. because it would never get off the ground. $\endgroup$ – a4android Nov 27 '20 at 5:59
  • $\begingroup$ @a4android Agreed, like I said in the post and in my comment, there's a lot of debate. But, if it did work, it's the best we have in that class (the only known reactionless drive!) and the start of a revolution - since as I said above it violates known laws of physics. Even a microscopic amount of thrust is amazing when it violates laws of physics and doesn't require mass! It's a great example for OP since it's the same situation (something we can build but don't understand) and it even involves space craft! $\endgroup$ – gbeeduljqa Nov 27 '20 at 7:04

The easiest way to accomplish this is to create a replicator. Require use of one of the artifacts during the process of assembling a new one. In this step, the device does something, which proves quite essential, even if nobody knows what it is doing.

One might draw a corollary to reproduction, actually. People did it for quite a long period of time without understanding what was truly happening. One might argue we still don't understand it, although we're getting a modicum of understanding of how the first 4 or 5 cell divisions work.


I don't know that this is worthy of a full answer of its own, particularly as you have already marked a solution, and I am sure it is a thing you have thought about already...


The key stumbling block in all the answers posted so far is "impervious". This could mean one of two things:

  1. Figuratively: The key to understanding the technology properly is several jumps away and we aren't there yet... but we could get there with more time and research. This former seems to be possible, and the only quibbles are about how long it might take us and imagining a tech a long way beyond what we know now.

  2. Literally: The key to understanding the technology is unattainable to us, even though we are able to use it at black-box level. This is more challenging, in that we would have to admit that there are certain things we could never know... and giving in is not a human trait.

To attempt a literal scenario 2, how about something that requires so much prior knowledge and understanding that we are physically incapable of storing it in our brains. Alternatively, it could be stored in a human brain but it would take so long to learn the details that it would take more than a human lifetime to comprehend?

To counter this, perhaps the aliens have extra biotech that allows them more space to know and compute things than we do, or their brains are just much more efficient at data storage and retrieval (through nature or meddling). Or perhaps they just live a lot longer than we do.

Again, neither of these necessarily put the technology permanently out of reach... Depending on what technology may come to help humanity. But even if we did get around it, the solution would be technology to help us with the learning/understanding process, not a stepping stone technology to the hyperdrive. And this does kind of meet the requirements.

  • $\begingroup$ Oh, no, I was talking figuratively. It's that just those "several jumps" are more like a couple of hundred millions of years of scientific advancement, so functionally it is impervious, I think - the sheer amount of "catching up" you'd need to do makes it so. $\endgroup$ – Darth Biomech Nov 27 '20 at 5:10
  • $\begingroup$ I think the main thrust of my response is that, while we don't want to admit it, human beings do have inherent physical limitations that may be different in an alien species. Also that these may provide you the required difficulty jump without going to the "could we ever get there" question. i.e. it isn't about whether humans can learn, it's about human capacity. $\endgroup$ – Graylocke Dec 1 '20 at 5:59
  • $\begingroup$ Thankfully, we have transhumanism. $\endgroup$ – Darth Biomech Dec 1 '20 at 10:14
  • $\begingroup$ Yes! Well some of us are more thankful than others, perhaps :P $\endgroup$ – Graylocke Dec 2 '20 at 3:15


Understanding isn't required on every level. Someone can make a watch with a schematic and the parts without ever understanding the battery.

Even better. When they made the electron microscope they didn't understand how small they were looking at first. They researched something that they knew had smaller wavelengths than light, allowing for higher resolution pictures, but the full scope was only understood later with further advancements of science.

Other ways to replicate without understanding is just not having the right equipment to understand what is happening. Our current theories (mostly) suggest that you can't go faster than light. So something here is happening that is definitely outside the scope of our current science. Much like the watch example you can set all the pieces in the right place, while not understanding or being able to measure the battery as you only have a microscope. They just don't have the right measuring tools. But they can just add material A there and B there, add a current and see it move.

  • $\begingroup$ Your first sentence doesn't make sense to me. The battery isn't something laying around that the lower-tech society can grab off a shelf and put into the watch they just assembled. They'd have to duplicate that battery, too. Are you assuming that it's OK with the OP that people use their own tech to re-create the object? As in, "we know that's a battery but don't know how to duplicate it, but we have batteries that would work, they'd just need to be carried in a backpack to use the watch..."? $\endgroup$ – JBH Nov 26 '20 at 20:54
  • $\begingroup$ @JBH it's just an introduction to the topic. It's not meant as a full metaphor for the whole thing. Full uderstanding isn't required in all cases. $\endgroup$ – Trioxidane Nov 26 '20 at 21:02
  • $\begingroup$ I'm an electrical engineer and the idea that a society could replicate an object but not understand the object it replicated is senseless. Now, I'll grant, maybe at first (I just wrote an answer that pointed this out), but once the we-never-knew-you-could-do-that! battery was replicated, people would instantly be working on understanding it. The question posits that they would fail.... From an engineering perspective it's like saying "I can build a transistor but don't know why it works." It's an impossible level of ignorance. $\endgroup$ – JBH Nov 26 '20 at 21:16
  • $\begingroup$ @JBH your understanding is moot. You can explain how to make a dynamo and a resistor for heating. Someone else might be able to explain how to create each component. Now imagine leaving these items and blueprints for other aliens. Intelligence and understanding are flexible both in concept as in the result. The aliens might be perfectly capable of replicating the process, but their different intelligence might simply prevent understanding. Aliens could give us the materials and the means for FTL, while our intelligence can simply not comprehend it (yet). $\endgroup$ – Trioxidane Nov 26 '20 at 22:02
  • $\begingroup$ "But their different intelligence might simply prevent understanding." That bought you a -1. That asserts that they can't experiment, test, or think their way out of the problem - despite having already proven that they can do so by creating the means for replication in the first place. $\endgroup$ – JBH Nov 26 '20 at 22:10


While similar to a subset of this answer I give you a worked example in the Chip Fab

Here they use lithography to apply different chemicals in patterns on a piece of silicon in layers.

There are several layers of abstraction that have gone on in this process.

  1. The "program" was designed.
  2. the logic to implement that was derived.
  3. the logic gates and interconnections to implement that were derived
  4. a layout for those logic gates and traces was derived
  5. the actual doping pools to implement the logic gates and traces was derived.
  6. masks to facilitate the doping were generated
  7. the chip was made

The people who work at each layer don't have to know how the other layers were derived, and none of them have to understand the quantum tunneling that makes semiconductors work.

To replicate a chip, you have to understand that there are layers of chemically treated silicon, and know what those chemicals are, and you have to know how to lay it out - in effect you only need to be able to understand steps 6 and 7 above.

To find out how to lay it out, all you need to do is take the chip apart and copy what you can see.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.